ishu-005/Algerian-Forest-Fire-FWI-Predictor

A Flask web app that predicts the Fire Weather Index (FWI) for the Algerian Forest Fires dataset using a trained Ridge regression model. The app loads a saved scaler and ridge model and returns a numeric FWI prediction from user-provided weather/fire-index inputs. Algerian Forest Fire — FWI Predictor A Flask web app that predicts the Fire Weather Index (FWI) for the Algerian Forest Fires dataset using a trained Ridge regression model. The app loads a saved scaler and ridge model (models/scaler.pkle, models/ridge.pkle) and returns a numeric FWI prediction from user-provided weather/fire-index inputs. What this app predicts Target variable: FWI (Fire Weather Index). The notebooks show y = df['FWI'] and model training used multiple regression methods and saved a ridge model. Inputs (form fields expected by the web UI) When calling the prediction route, the app expects these numeric inputs (order used by the scaler/model): Temperature RH (Relative Humidity) WS / Ws (Wind Speed) Rain FFMC DMC ISI Classes (encoded: not fire → 0, fire → 1 in preprocessing) Region (numeric region code) The app transforms input with standard_scaler.transform([[Temperature,RH,WS,Rain,FFMC,DMC,ISI,Classes,Region]]) and runs ridge_model.predict(...). Project structure ├── application.py # Flask app (routes: / and /predictdata) ├── requirement.txt # Python dependencies ├── dataset/ │ ├── Algerian_forest_fires_cleaned_dataset.csv │ └── Algerian_forest_fires_dataset_UPDATE.csv ├── models/ │ ├── ridge.pkle # trained Ridge regression model │ └── scaler.pkle # StandardScaler used for preprocessing ├── notebook/ │ ├── algerianforestfires.ipynb # EDA and cleaning │ └── modelTraning.ipynb # model training & evaluation (train/test split, model selection) └── templates/ ├── home.html └── index.html